We recently read an interesting article by Linxweiler et al, who reported the use of p16-Ki67 immunocytochemistry combined with human papillomavirus (HPV) testing on cytological specimens for the assessment of HPV involvement in head and neck squamous cell carcinoma (HNSCC).1 Their findings regarding brush cytology and HPV testing are in agreement with the results of our previous study.2 In fact, we recently reported that abnormal brush cytology is strongly associated with a diagnosis of HNSCC and that HPV testing on cytobrushing specimens from patients with HNSCC is feasible and reliable.Linxweiler et al found a very high sensitivity (98%) and specificity (100%) for head and neck brush cytology, although the estimation of sensitivity based on an analytic population that was composed mostly of patients with cancer (50 of 65 patients) has to be regarded with caution.1 In fact, a recent work highlighted that disease prevalence affects test characteristics, including sensitivity.3 In addition, the high sensitivity reported by Linxweiler et al may depend on the fact that the samples were partly obtained during a panendoscopy, a procedure that guarantees a more thorough collection of cytological specimens compared with an ambulatory setting. Indeed, in an ambulatory setting, we collected 100 cytobrushing specimens from patients with HNSCC, and detected abnormal cells in only 77 of them (77%).Linxweiler et al also claimed that false-negative results regarding HPV status assessed on the cytological samples may be avoided by screening the samples for the presence of cancer cells prior to HPV testing. 1 We performed the HPV test on 68 paired cytological-histological specimens from patients with HNSCC (Cohen kappa, 0.84) and based on our results, we only partly agree with the authors. In fact, among 9 morphologically negative cytobrushing specimens that corresponded to HPVpositive tumors, 3 specimens tested negative for HPV (thus representing false-negative results), whereas the other 6 were positive for HPV type 16, a finding that is in agreement with the respective tumor tissue. These findings demonstrate that cytological brushings may yield a reliable HPV test result despite the fact that they do not contain neoplastic cells.Although the performance of head and neck cytology can only be evaluated on a population-based sample, we believe that the study by Linxweiler et al represents a strong contribution toward promoting the use of brush cytology in patients with HNSCC for the assessment of HPV involvement in such tumors. We thank Drs. Dona and Benevolo for their constructive comments on our recent article, 1 which pointed out some aspects to be discussed further. When comparing the diagnostic sensitivity of liquid-based cytology to screen for squamous cell carcinoma of the head and neck region
382Cancer Cytopathology